[btoptical]

Really none of that is correct.

Firstly it's not at all equivalent to an AM radio. It is an AM modulated carrier but that's about it. The detection scheme direct detection and not coherent like in an AM radio.

Also your power spectral density would be fairly low as your AM modulation would be distributed over a huge wavelength range. So taking a narrow slice with a tunable optical filter (which aren't that narrow really) would mean the detector would see a very small signal indeed.

Given that it's so broadband, my guess is that it's probably not low intensity noise at all (modern communications lasers are ~-140dB/Hz) and thus wouldn't make a great direct detection OOK source. So this would limit ultimately the sensitivity of such an AM modulated system. Being so broadband means that it's not at all suitable for phase modulated signal because by definition it would have high phase noise. Overall the spectral efficiency of such a source as a communications device is abysmal.

Additionally the silicon detectors that are needed to convert the light signal back into an electrical signal have low responsivity and speed. So the poor detector speed places an upper bound on the supportable BW and poor Si responsivity places a limit on the overall sensitivity of the system.

It's much, much more efficient to use IR sources all the way around. Materials in the IR are really efficient for both light generation and detection. There are lots of other reasons why IR is better for free-space communications as well (less scattering and higher material transparencies generally).

No a white laser is good for other reasons. There are lots of places where we still use things like Xenon lamps for measurement of things.